Joint MMSE Transceiver Designs for MIMO AF Relaying Systems with Direct Link

In this paper, we provide minimum mean-squared error-based source-relay-destination transceiver designs for multiple-input multiple-output amplify-and-forward relaying systems, where direct link between the source and the destination is non-negligible. In an earlier work, a local optimal technique was introduced which employs a projected gradient method and an interior point method. Since these methods may have quite high computational complexity, we investigate a new local optimal solution for the source-relay-destination transceiver which has low complexity. To this end, we first introduce the optimal closed-form solution for the relay transceiver for given source and destination filters. Then, for given relay and destination transceivers, the optimal source precoder design is derived, which requires only 1-D bisection search. Based on these solutions, we propose a joint optimization algorithm which iteratively finds a local optimal solution. Also, we introduce a simple non-iterative algorithm which computes filters in closed-forms with low complexity. Furthermore, since perfect channel knowledge may not be feasible in practical systems, a joint transceiver technique which is robust to channel uncertainties is provided. It is confirmed by simulation results that the proposed schemes outperform conventional techniques with significantly reduced complexity.